What causes length of stay outliers?
LOS outliers — patients staying 2x or more the expected geometric mean — are driven by three categories: clinical complications requiring extended treatment (35%), discharge disposition delays waiting for post-acute placement (40%), and social factors including insurance authorization for SNF or home health (25%). The disposition delays are the most actionable category because they represent clinically-ready patients held by administrative and logistical barriers.
What this looks like in Vizier
Stylized dashboard visualization. Data values obscured. Upload your own data to see real numbers.
Why This Happens
Clinical complication outliers arise from wound infections, pulmonary embolism, hospital-acquired delirium, and in-hospital falls — each adding three to seven days beyond the expected LOS. These events are partially preventable through bundle compliance (CAUTI, CLABSI, VTE prophylaxis) and fall prevention programs, but they cannot be eliminated entirely in a complex inpatient population. They account for roughly 35% of LOS outlier events and carry the highest clinical acuity among the three categories.
Disposition delay outliers — the largest category at 40% — represent a fundamentally different problem. These patients are clinically ready for discharge but are waiting for SNF placement, home health authorization, or durable medical equipment delivery. SNF placement requires insurance authorization that averages two to four business days, followed by SNF bed availability that averages an additional 1.3 days. Most hospitals do not initiate the authorization process until the day the physician determines the patient is clinically ready, adding three to five preventable inpatient days per event. Social factor outliers (25%) involve patients with no safe home environment, active domestic violence situations, or homelessness — situations requiring social work intervention that is frequently triggered on day three of admission rather than at the time of admission screening.
What the Data Usually Hides
LOS outlier reporting typically surfaces the distribution of outlier patients by DRG. A hospital might see 40 LOS outliers clustered in DRGs 291 (heart failure with MCC) and 470 (major joint replacement) and assume these represent clinical complexity. In reality, reviewing the reason for outlier status would reveal that 35 of those 40 patients were disposition delays or social factor holds — not clinical complications. The intervention priority is completely reversed by this misclassification.
Excess day cost calculations typically use the hospital's fully-allocated cost per day. This overstates the cost of clinical outliers (which consume real resources) and understates the cost of disposition delay outliers (which consume bed-days that could serve other patients). A more precise analysis separates variable cost per outlier day from the opportunity cost of the blocked bed, which at peak census periods can exceed the direct care cost.
How to Fix It
Implement LOS outlier reason coding at discharge — a required field in the discharge summary that categorizes each outlier as clinical, disposition, or social. This single data element transforms a volume metric into an actionable segmented report. The coding burden is low (three categories, documented at discharge) but the analytical value is high: it enables root cause separation that drives different interventions for each category.
For disposition delays, the intervention is a day-two trigger: for any patient whose anticipated discharge is beyond three days, social work and case management begin SNF pre-authorization on day two of admission. For patients with complex social situations, admission-day social work screening (not day-three screening) identifies barriers before they become three-day delays. SNF pre-authorization workflows that run concurrently with clinical care — rather than sequentially after clinical readiness — reduce disposition-delay LOS outliers by 40–60% in hospitals that have implemented them systematically.
Your Data. Your Answer.
This is what the data typically shows.
Want to see what your data says?
Ask Your Vizier →